Peer Revenue Past citations are no guarantee of future returns

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Sunday 13 September

Distributed Computing

Flattening Every Memory Peak in Long-Context Mixture-of-Experts Training

Training a Mixture-of-Experts (MoE) model at long context or large batch size fails as soon as any one component's peak allocation exceeds device memory, so the target is every peak at once, not the average footprint. Four are left unbounded by the parallelism plans in common use, and each grows differently: expert dispatch with the routing matrix, the vocabulary projection with tokens times vocabulary, gradient checkpoint boundaries with depth times sequence length, and optimizer state with parameter count. Which one runs out first changes with the model, the context length, and the device count, so lowering the largest only exposes the next. We bound all four with schedules whose GPU working set is fixed at launch: PipelinedLLEP extends least-loaded expert parallelism with a cap on the tokens each source contributes to a dispatch chunk, Ring-DTP circulates activations or weight shards around a ring at the vocabulary projection and folds each block of logits into an online log-sum-exp, Selective checkpoint offload (SCO) keeps the one long-lived tensor of each checkpoint boundary in CPU memory, and OffloadStreamAdamW turns the serial CPU Adam update of optimizer offload into a bucket pipeline. All four change only the order and granularity of computation and data movement, so the loss and gradients stay exact. In matched component tests, they cut the MoE dispatch peak by up to $59.3\%$ without losing throughput, the vocabulary projection peak by $86.6\%$, and the offloaded optimizer step by $2.05\times$ faster. Composed on MoE models from 120B to 667B parameters, they train at 1M context length, $8$–$32\times$ the reach of a tuned FSDP2 baseline, and up to $10.4\times$ its throughput.
Read the full abstract

All authors Shrey Pandit, Xuan-Phi Nguyen, Yiran Zhao, Shafiq Joty

High Energy Physics (Theory)

Pseudo entropy from entanglement entropy

Pseudo entropy extends entanglement entropy from a single quantum state to a pair of nonorthogonal states and is generally complex. Taking advantage of Cauchy-Riemann equations, Kramers-Kronig relations, and analytic continuation in state parameters, we show how and to what extent real and imaginary parts of pseudo entropy can be derived from ordinary entanglement entropy. For families with holomorphic coefficients in finite-dimensional Hilbert spaces, the reduced transition matrix equals the ordinary reduced density matrix formula evaluated at complex parameters. Then a convergent Taylor series gives the real and imaginary parts of pseudo entropy from even and odd derivatives of entanglement entropy at the real midpoint. The matrix identity also gives formulae for excess pseudo entropy and Renyi entropies, and interpolation formulae for families with polynomial coefficients. We apply these results to boundary-state quenches and thermal states in conformal field theory, and to fermionic and bosonic Gaussian states and quenches. In conformal field theory, Kramers-Kronig relations give the first moment of the imaginary part in terms of the central charge and one-point functions for boundary states.
Read the full abstract

Analysis of PDEs

On the stability of solutions to the 3-D anisotropic Navier-Stokes equations in the critical space

In this paper, we study the continuous dependence of solutions to the three-dimensional incompressible anisotropic Navier-Stokes equations $(ANS)$ with initial data in the critical space $\dot{B}^{0,\frac12}$. We prove that the data-to-solution map is continuous in this setting. Due to the critical regularity of the problem, the classical Bona-Smith argument \cite{BoSm} (see also the abstract result \cite{ABITZ}) does not appear to apply because a key ingredient (the Lipschitz continuity at lower regularity) does not appear to hold. As a consequence, we develop a new direct Fourier-weighted method within the Besov framework.
Read the full abstract

Numerical Analysis

Regional Stability and New Eigenvalue Perturbation Bounds

Let $A$ be an $n\times n$ symmetric matrix with eigenvalues $\lambda_1\geq\cdots\geq\lambda_n$. Let $\tilde A:=A+E$, where $E$ is a symmetric noise matrix, and denote the eigenvalues of $\tilde A$ by $\tilde\lambda_1\geq\cdots\geq\tilde\lambda_n$. Bounding the perturbation $|\tilde \lambda_i -\lambda_i| $ is a central problem in linear algebra and numerical analysis. In this paper, we prove new perturbation results by exploring the actual interaction between $E$ and the eigenvectors of $A$. In the setting where $E$ does not act adversarially with respect to these vectors (for instance, if $E$ is random), we obtain a considerable improvement over Weyl's inequality. One can routinely extend these results to the Hermitian and rectangular settings. We obtain the new bounds as corollaries of a regional stability result, which provides a sufficient condition for a region on the real line to be stable (containing the same number of eigenvalues) after the perturbation. This result is of independent interest. We prove our regional stability result using contour integral analysis. Our main new technical ingredient here is the double-jump argument, which is robust and could be useful in many other situations involving Neumann series.
Read the full abstract

Computation and Language

E2A-Bench: Benchmarking Evidence-to-Action Reliability in Financial Chart Reasoning

Can financial vision-language models (VLMs) turn chart evidence into reliable action recommendations? Existing hallucination evaluations are mostly claim-centric; they assess whether generated statements are supported, but not whether evidence remains traceable through rationale, confidence, and final action. We introduce E2A-Bench, a 969-query benchmark for financial chart reasoning, constructed from 323 HS300 constituents under three input modalities with deterministic OHLCV-derived evidence anchors. E2A-Bench evaluates grounding, reasoning-action consistency, evidence-confidence calibration, and directional coverage through UCR, RCI, ECI, and NDR, where NDR measures coverage-aware evidence-to-action reliability rather than realized trading performance. Evaluating 20 VLMs reveals three failures hidden by scalar hallucination scores: the lowest-UCR model ranks near the bottom by NDR due to only 6.4% directional coverage; oracle-aided verification reduces unsupported claims but can collapse coverage; and financial fine-tuning amplifies the BUY:SELL ratio by factors of 4.21 to 4.68 across strict base-fine-tuned pairs. These results show that financial VLM evaluation should trace the full evidence-to-action chain rather than rely on a single hallucination score. Code and data: https://github.com/wanng-ide/E2A-Bench
Read the full abstract

Game Theory

Approximating Optimal Welfare in Complementary Allocation under Decentralized Information

Complementary resources are often allocated by agencies that see different coordinates of an individual's needs. If an intervention requires complementary resources, an agency may know if said person lacks its own resource; yet not know if supplying completes a useful bundle. We study allocation of $m$ divisible complementary goods when each agency observes only its own coordinate of a baseline-access profile. To isolate information from incentives, we measure the welfare cost of this split against a generous decentralized benchmark DEC: the best rule that acts on local information alone, with cooperative agencies and known population distribution and capacities. Even so, local observation alone can make optimally coordinated policies arbitrarily inefficient: a factor linear in $m$ (even with equal capacities), and a factor inverse in OPT (with but three agencies). Threshold referrals recover much of this loss: an agency reports if its value lies below a public threshold, and a clearing rule uses only the joint reports. Aggregating types by their reports captures a rectangle under the optimal half-utility survival curve. Our main tool is a profile-level charging certificate: if optimal utility is at least twice a threshold, OPT pays at least that on every deficient coordinate. This yields welfare at least ${\rm OPT}/[4(1+\ln(2/{\rm OPT}))]$ with one bit per agency; and an equal-revenue family shows this log-loss tight for one threshold. A geometric threshold ladder improves this to a constant fraction of OPT with doubly-logarithmic messages: a staircase recovers ${\rm OPT}/8$ with $\Theta(\log\log(1/{\rm OPT}))$ bits, and this is necessary within the class. The results isolate how a simple reporting language turns large local-info losses into constant-factor recovery. In short: One threshold gets a log-approximation. A staircase, with log-of-log bits, a constant-factor.
Read the full abstract

Number Theory

Cubic AGM graphs and Hessian $3$-isogenies over finite fields $\mathbb{F}_q$ of odd characteristic with $q \equiv 2 \pmod3$

Let $\mathbb{F}_q$ be a finite field of odd characteristic with $q \equiv 2 \pmod3$. We study the directed graph defined by the Borwein–Borwein cubic arithmetic–geometric mean (AGM) over $\mathbb{F}_q$. We prove that this graph is a disjoint union of directed cycles. We associate Hessian curves with this AGM. We also show that each edge corresponds to a $3$-isogeny defined over $\mathbb{F}_q$ between the curves associated with its initial and terminal vertices. We then use a counting formula for Hessian curves to derive a lower bound for the number of cycles.
Read the full abstract

Machine Learning

Learning Source Acquisition Policies by Offline Planning

Predicting under an acquisition budget requires choosing feature groups whose value can depend on later queries. O-MPAC transfers finite-horizon risk-cost targets from complete training records into a shared source-action scorer. At inference time, the scorer uses partial observations and source metadata, re-scores after each query, and applies a hard cost mask. We analyze how tied teacher targets and the remaining planning horizon affect the learned decisions. Uniform supervision over tied minima preserves the target distribution under source relabeling. In a five-seed routing experiment, it achieves 0.965 accuracy under both original and context-last orders. On six real tasks, validation selects H1 without action cross-entropy in all thirty splits. O-MPAC has the highest mean budget-integrated accuracy on five tasks against source-adapted GDFS, DIME, AACO+NN and a static policy.
Read the full abstract

Classical Analysis and ODEs

Non-radial Dunkl multipliers at the $L^2$-Sobolev threshold

We prove a Hörmander multiplier theorem for the Dunkl transform associated with an arbitrary finite reflection group. Uniform $H^\sigma(\mathbb{R}^N)$ bounds for the normalised dyadic pieces of a measurable symbol, with $\sigma>\mathbf N/2$ and $\mathbf N$ the homogeneous dimension, imply $L^p(d\omega)$ boundedness for $1<p<\infty$ and weak type $(1,1)$. No radiality or reflection-group invariance is assumed, and the dyadic pieces may meet the reflecting hyperplanes. The main new estimate controls the Dunkl kernel $E_\kappa(i\xi,x)$ and its first Euclidean derivatives in $x$, uniformly near arbitrary intersections of reflecting hyperplanes. Here $\xi$ lies in a fixed compact annulus. Writing $w_\kappa$ for the Dunkl weight and $\Omega_\kappa$ for its inhomogeneous counterpart, it gives the bound $C\Omega_\kappa(x)^{-1/2}w_\kappa(\xi)^{-1/2}$. The proof decomposes each root contribution into Hermitian two-dimensional blocks, applies oscillatory changes of variables near the rootwise transition points, and uses a commutator identity to cancel the non-integrable phase derivative. Chamber lifting also yields endpoint bounds for the pullbacks of the componentwise Coifman–Weiss atomic $H^1$ space and the corresponding BMO space on a fixed chamber.
Read the full abstract

All authors Der-Chen Chang, Ji Li, Chaojie Wen, Liangchuan Wu

Robotics

Multi-Task Visual Perception Network with LLM Conditioning for Autonomous Navigation

Long-term navigation for service robots faces crit- ical challenges like the accumulation of odometry drift and sensor error, which progressively degrade 2D maps and renders traditional path planning algorithms (e.g., A*, RRT*, DiPPer, ViT-A*) ineffective over time. To address this, we propose a user-friendly, interactive framework that eliminates the reliance on globally consistent maps. Our approach integrates visual perception with Large Language Models (LLM) to interpret user commands via text or voice. Instead of relying on a drift- prone global map, the system generates a sequential action plan based on local visual cues and egocentric geometric instructions. These action plans are executed sequentially, allowing the robot to navigate known and unknown environments safely. By reset- ting localization relative to immediate targets, our framework effectively works with a minimum accumulation drift strategy, ensuring accurate, efficient, and collision-free navigation without the maintenance overhead of traditional mapping. Experiments on real-world and simulated data have shown significant improve- ments over other methods. Our source code is publicly accessible at https://github.com/PraveenSingh24/VL-Navigation.
Read the full abstract

High Energy Physics (Theory)

Holographic Casimir Effect for Mixed Boundary Conditions and non-CFTs

This paper investigates the holographic Casimir effect in AdS/BCFT with a brane-localized scalar field. The scalar field takes different values at the strip's two boundaries, leading to mixed boundary conditions. The massive brane-localized scalar field typically breaks the boundary conformal symmetries, allowing us to study the holographic Casimir effect in non-CFTs. In two-dimensional spacetime, we apply the holographic g-theorem to prove the holographic bound on the Casimir effect for general brane-localized matter fields. In higher dimensions, we find that the brane-localized scalar field reduces the Casimir amplitude when the mass squared \(m^2=0\), but can increase it when \(m^2 < 0\). Additionally, there is a no-hair theorem for the cases where \(m^2>0\). These findings indicate that relevant boundary deformations can enhance the Casimir effect. As a byproduct, we obtain a gravitational dual of the repulsive Casimir force under mixed boundary conditions and argue that it aligns with the cosmic censorship conjecture.
Read the full abstract

Number Theory

The Jensen–Pólya program and inequalities for arithmetic sequences

We develop an analytic framework to study hyperbolicity of Jensen polynomials and related inequalities for arithmetic sequences with general asymptotic growth. Our results provide partial converses to the Hermite–Jensen phenomenon of Griffin, Ono, Zagier, and the third author. As applications we prove several conjectures concerning log-concavity, higher-order Turan inequalities, Laguerre inequalities, infinite log-concavity, and Toeplitz determinants for partition-type functions.
Read the full abstract

Materials Science

Time- and Frequency-Resolved Observation of Inverse Orbital Hall Effect in Gallium Nitride via Terahertz Polarimetry

Orbitronics has attracted significant attention as a platform for information storage and processing that exploits the orbital angular momentum (OAM) of electrons without the need for spin-orbit coupling. However, experimental evaluation of OAM-charge interconversion remains a challenge and the results are often controversial due to the coexistence of bulk and interfacial contributions and the complexity in sample structures. Here, using circularly polarized light pulses and terahertz (THz) polarimetry, we developed a non-contact method to observe the inverse orbital Hall effect as a bulk response within a single material without employing heterostructure samples. In a semiconductor GaN, we directly captured the OAM-to-charge current conversion of holes in the THz frequency range. By analyzing the sharp frequency dependence of the Hall conductivity, we disentangled the competing microscopic mechanisms and revealed the dominant role of extrinsic contributions to the orbital Hall effect in the dc limit. By contrast, the Hall conductivity at THz frequencies above the impurity scattering rate is attributed to the intrinsic Berry-curvature mechanism,allowing a quantitative comparison with the microscopic theory. Furthermore, the ultrafast dynamics of the inverse orbital Hall signal directly revealed sub-picosecond OAM relaxation of holes, comparable to that of phonon-mediated thermalization. The quantitative argument based on theoretical calculations suggested the existence of an even faster decay channel due to momentum redistribution by acoustic phonons, suggesting a sub-nanometer-scale OAM relaxation length. Our results provide comprehensive and crucial insights into OAM transport and establish an ultrafast, contact-free approach for investigating OAM-to-charge conversion.
Read the full abstract

All authors Kota Aikyo, Tomohiro Fujimoto, Ami Mi Shirai, Yuta Murotani, Mitsuru Funato, Shinji Miwa, Jun Yoshinobu, Yoichi Kawakami, Ryusuke Matsunaga

Materials Science

Cryothermal Measurements of Variable-Emittance Coatings with Lower Phase Transition Temperatures for Space Thermal Control

Space thermal control is critically important to ensure proper operation of on-board equipment in a regulated temperature range. Passive thermal control with variable-emittance coatings (VECs) could help save power consumption in a dynamically changing space thermal environment. Vanadium dioxide (VO2) based VECs have been studied for space thermal control but its intrinsic phase transition around 68{\deg}C limits its wider space applications where lower temperature ranges are expected. In this work, we experimentally demonstrate enhanced radiative heat dissipation in space-like thermal environment via cryothermal measurements with VECs of undoped and tungsten doped VO2. The fabricated undoped VEC exhibits a large emittance change of 0.6 across the phase transition, while the 1 at.% tungsten doped one shows an appreciable emittance variable of 0.4 with phase transition temperature lowered by 25{\deg}C. A vacuum cryothermal setup is developed with a liquid nitrogen cooled coldfinger to mimic cold space thermal background and a custom-designed sample mount suspended by nylon wires. After careful calibration and validation, greatly enhanced radiative heat dissipation upon VO2 phase transition up to 3.5 times with transition temperature lowered by 25{\deg}C from 1 at.% tungsten doping is clearly observed from the cryothermal tests. In the actual space thermal environment, radiative heat flux could further increase across phase transition from 160 W/m2 to 650 W/m2 with undoped VO2 coating from 55{\deg}C to 80{\deg}C, and from 175 W/m2 to 493 W/m2 with 1 at.% tungsten doped VEC from 30{\deg}C to 55{\deg}C.
Read the full abstract

Fluid Dynamics

A compatibility–realization framework for singular Navier–Stokes flows: admissible families and non-rigid core mechanics

We introduce a compatibility–realization framework for constructing singular solutions to the forced three-dimensional incompressible Navier–Stokes equations. The framework separates exact balance and matching constraints, expressed in a chosen geometric representation, from the sufficient conditions of a particular completion scheme. Completed realizations are organized by observable fibers, distinguishing prescribed core kinematics from the surrounding force and transport mechanisms. Conditional on the source-dependent completion input specified in this paper, we obtain families with zero initial velocity, smooth compactly supported forcing, bounded kinetic energy and finite-time unbounded velocity. These include a reversal of leading viscous work at fixed axis motion and first velocity gradient, a signed transition through a stationary singular center under one fixed canonical pressure trace, and a connected stretching family crossing the threshold of zero leading direct viscous vorticity supply. Near a critical profile, the realizable trace data contain a neighborhood in an infinite-dimensional space of analytic axial-velocity and pressure perturbations, while the central first velocity gradient and the critical vorticity balance remain fixed. We also identify the full kernel and an explicit right inverse of the joint angular-transport operator in the chosen realization class. The results establish non-rigidity within observable fibers and show that axis kinematics alone do not determine the local pressure–viscosity partition. Direct profile selection and dynamical stress generation are treated as coupled components of construction, rather than as definitions of the entire design space.
Read the full abstract

Differential Geometry

Characterization of contact structures for the dual form of the vector field of an h-Ricci-Bourguignon soliton on the product manifold D2 and R

In this paper, we study h-Ricci-Bourguignon solitons on the product manifold D2 x R, where D2 is the Poincare disk equipped with its standard hyperbolic metric. We first establish that this manifold admits a gradient Ricci-Bourguignon h-soliton only if the non-zero function h depends solely on the manifold's last coordinate. Next, in the case where h:=1, we explicitly determine the vector field that makes D2 x R a Ricci-Bourguignon soliton. Finally, we provide the necessary and sufficient condition for the dual 1-form of this vector field to define a contact structure, which corresponds to the existence of an explicitly determined function independent of the manifold's second local coordinate, thereby allowing us to compute its Reeb vector field.
Read the full abstract

Cryptography and Security

Fusing Spectral Signatures and Activation Clustering for Backdoor Detection in Healthcare Imaging Models: Method, Implementation, and Evaluation

Machine learning models are increasingly deployed in healthcare imaging pipelines for diagnostic support, and training-time attacks against them are a named sector-level concern: healthcare-sector guidance identifies model poisoning and adversarial attacks as threats requiring dedicated defenses, while federal policy directs expanded AI vulnerability-detection tooling to critical infrastructure operators such as rural hospitals. Spectral signature analysis and activation clustering are two established backdoor detection methods routinely evaluated as independent baselines, but their outputs are not ordinarily combined, and reported detection performance on medical imaging benchmarks remains sparse relative to the natural-image setting. This paper contributes three things: a score-level fusion rule combining per-class spectral ranking with activation-clustering flags into a single per-sample poisoning score and a model-level agreement statistic; an open-source implementation of the resulting eight-stage pipeline; and an evaluation of that pipeline against synthetically poisoned variants of a public medical imaging benchmark and CIFAR-10 at four poisoning rates (0%, 1%, 5%, 10%) over five seeds each, measuring each detector alone against the fusion. On the medical benchmark, the fused detector reaches AUROC >= 0.99 at every nonzero poisoning rate tested. On CIFAR-10, fusion does not uniformly help: at 10% poisoning, activation clustering's true-positive rate collapses to 0.000 and spectral AUROC independently degrades to near-chance (0.545), despite a 97.2% attack success rate confirming the backdoor was fully installed. The fused score, a weighted combination of both signals, inherits this joint failure. Detection output is expressed in NIST AI RMF Measure-function and MITRE ATLAS terms, so findings are reported in the vocabulary security and compliance teams already use.
Read the full abstract

Human-Computer Interaction

AnnoSketch: Evaluating and Collecting Human Sketches for MLLM-assisted Chart Annotation

As multimodal large language models (MLLMs) support a growing range of input modalities, increasing work explores how to incorporate rough sketches to convey user intent. For annotated chart generation, it remains unclear what annotation sketches people provide and when such visual input helps MLLMs generate more useful annotations. In this study, we examine when sketch input is useful for MLLM-generated chart annotations across variation in chart type and caption type. In addition, we qualitatively analyze participants' explanations of their output preferences to characterize what made generated annotations more or less helpful. To further document participants' annotation sketches, we present AnnoSketch, comprising 1,600 annotation sketches collected across 160 chart-caption pairs from the conditions in which sketch guidance proved most beneficial, together with participants' annotation intents, perceived comprehension difficulty, and self-reported expressive limitations. We also label these sketches with structured metadata describing how each sketch relates to its caption and how participants express annotations through visual marks. Together, our study and AnnoSketch help determine when to solicit sketch input and provide empirical source for how people sketch chart annotations to support captions. The dataset and supplemental materials are available in our OSF repository.
Read the full abstract

All authors Yoonjae Oh, Seon Gyeom Kim, Jae Young Choi, Ryan Rossi, Jihyung Kil, Eunyee Koh, Tak Yeon Lee

Computation and Language

Editorial routing shapes how computational results are qualified in AI-assisted scientific writing

Large language models increasingly analyze computational results and draft manuscripts, making reliable communication as important as correct analysis. Using fixed computational evidence, we tested whether assigning comparisons across modeling choices elsewhere in a research workflow changes manuscript reporting. In constrained sentence-writing tasks, Anthropic's Claude Sonnet 5 often omitted numerical qualifications when detailed comparisons were assigned to a group repository, but retained them more often when the same comparison was assigned to Supporting Information or its own working notes; Claude Opus 5 was less sensitive. These effects did not follow a simple accessibility ordering. A targeted placement rule largely restored sentence-level qualification, whereas a generic accuracy reminder did not. Longer contributions retained numerical qualifications, although some summaries across computational settings were still redirected to the repository. Thus, documenting context within an AI workflow does not ensure its communication where readers encounter the result.
Read the full abstract

Analysis of PDEs

Guaranteed wave-speed bounds for the compressible Euler equations with a composite equation of state

We derive computable upper bounds on the maximum wave speed in the Riemann problem for the compressible Euler equations with a composite equation of state, in which the pressure is the sum of a convex hydrodynamical part and a barotropic correction. The correction may destroy the convexity of the equation of state, so that the Riemann solution can contain composite shock-rarefaction waves. The bounds are obtained by comparison with the auxiliary Riemann problem for the hydrodynamical equation of state alone. The bounds hold irrespective of whether the outer waves are elementary or composite and require no case distinction on the wave type.
Read the full abstract

Cryptography and Security

Policy-Governed Post-Quantum Migration for Legacy Microservices Using Ephemeral Sidecar Architectures

The fast development of quantum computing represents a big risk to classical cryptography that is commonly used in cloud native and microservice based enterprise systems. Traditional cryptographic primitives are closely linked to legacy microservices and it is both intricate, hazardous, and disruptive to straight up migrate to post-quantum cryptography (PQC). In a bid to overcome these issues, this research presents a PolicyGoverned Post-Quantum Migration through Ephemeral Sidecar Architectures (PG-PQMES), a framework of dynamic and reversible migration that allows transparent adoption of PQC without any modifications in the legacy code of applications. The architecture is a combination of three coherent layers, including Ephemeral Crypto Sidecar Layer which injects the runtime cryptography, Policy Governance Layer which manages the migration centrally, and Migration Safety and Observability Layer which measures the performance and rolls back automatically. An innovative Policy-Governed Ephemeral PQ Migration (PG-EPM) algorithm is proposed to maximize the performance, compliance, and trust-based migration. In simulated environment of microservices, experimental assessment shows that the time of migration, service downtime, overheads of latency, and rollback recovery time are substantially reduced using the current migration strategies. The findings show that a sidecar-based migration strategy that is policy-based offers a viable, scalable, and enterprise-scale migration to a secure post-quantum transformation.
Read the full abstract

Category Theory

Characterizing (Co)Free Dagger Categories

For any category, there exists both a free dagger category and a cofree dagger category over it. A natural question to ask is: given a dagger category, how can we tell if it is free or cofree without specifying an external base category? In this paper, we provide characterizations of both free dagger categories and cofree dagger categories via internal dagger category structure. To characterize cofree dagger categories, we use rectangular bands and show that a dagger category is cofree if and only if it is enriched over rectangular bands. For free dagger categories, we define the notion of a zigzag dagger category, and then show that a dagger category is free if and only if it is a zigzag dagger category. We also show that free dagger categories can be characterized as the coalgebras of the induced comonad from the free dagger category adjunction, and similarly that cofree free dagger categories can be characterized as the algebras of the induced monad from the cofree dagger category adjunction.
Read the full abstract

Computer Vision

Vision-Language Models for Criterion-Level Grading of Handwritten Examinations in Outcome-Based Education

Criterion-level grading connects examination performance to learning outcomes, but manual marking introduces workload and variation between markers. This study evaluates vision-language models (VLMs) for handwritten outcome-based assessment across five dimensions: accuracy, human agreement, repeated-run reliability, error concentration, and explanation quality. Using 1,982 criterion-level records from 485 undergraduate examination answers, we compare 20 configurations spanning Qwen2.5-VL, InternVL3, Pixtral, a Donut baseline, and a cascade ensemble. Evaluation setups include zero-shot prompting, few-shot prompting, partial fine-tuning, and Low-Rank Adaptation (LoRA). Two independent faculty markers regraded all 291 test criteria, providing a human agreement baseline on the same assessment materials. Qwen2.5-VL with LoRA achieved Quadratic Weighted Kappa (QWK) of 0.727 and mean absolute error of 0.435 marks against the examiner, compared with mean human-pair QWK of 0.551. This comparison reflects calibration to the examiner's training marks. LoRA outperformed partial fine-tuning for all three instruction-tuned VLMs, while few-shot prompting reduced QWK in every configuration with valid prompted scores. Aggregate reliability and exact repeatability diverged: intraclass correlations ranged from 0.790 to 0.874, yet 50.2-63.6% of criteria changed marks across five sampled runs. Attention-guided deletion showed no statistically significant advantage over random masking, and four faculty reviewers reached no consensus on explanation usefulness. These findings highlight the need for rubric-specific calibration, repeatable scoring, review of consequential errors, and separate validation of explanations. The released evaluation protocol supports criterion-level assessment research and grading tools with teacher oversight.
Read the full abstract

All authors Asif Hasan Tonmoy, Saad Ahmed, Md Khalid Syfullah, S. M. Jahangir Alam

Algebraic Geometry

Torelli theorems for Quot schemes of vector bundles on curves

For a vector bundle $E$ on a smooth projective curve $C$, one defines the Quot scheme $Q_d(E,C)$ parametrizing subsheaves of $E$ having length $d$ torsion quotients. If $E_i$ is a vector bundle of rank $r\geq 2$ on a smooth projective curve $C_i$ for $i=1,2$, isomorphisms between $\mathbb{P}_{C_i}(E_i)$ preserving the projective bundle structures induce isomorphisms of $Q_d(E_i,C_i)$, called natural isomorphisms. We show that for $d\geq 2$ all isomorphisms between $Q_d(E_i,C_i)$ are natural, except for a non-natural involution of $Q_2(\mathcal{O}_C^{\oplus r},C)$. This gives a complete description of the automorphism group of $Q_d(E,C)$, generalizing previous works of Biswas-Dhillon-Hurtubise and Gangopadhyay. This also shows that one can reconstruct the curve from the Quot scheme, a Torelli-type theorem. As a key step in our proof, we show that any isomorphism between symmetric powers of smooth projective curves preserving big diagonals is natural, which is interesting in its own right.
Read the full abstract

Operator Algebras

The UMD property of symmetric operator spaces

We prove that if $E$ is a UMD symmetric Banach function space on $(0,\infty)$, then $E(\mathcal{M},\tau)$ is UMD for every semifinite von Neumann algebra $\mathcal{M}$ equipped with a faithful normal semifinite trace $\tau$. This resolves in the affirmative an open problem that has circulated in the non-commutative world for more than four decades.
Read the full abstract

Combinatorics

Tuza's Ryser-conjecture claim for four-partite hypergraphs with matching number two

We prove that every $4$-partite $4$-uniform hypergraph $H$ with matching number $\nu(H)=2$ satisfies $\tau(H)\le 6$, where $\tau$ denotes the vertex-cover number. This confirms a claim made by Tuza in his 1979 manuscript but never published with a proof, and closes the case $(r,\nu)=(4,2)$ of Ryser's conjecture. The best previous bound was $\tau\le 7$, an integrality consequence of the theorem of Haxell and Scott (2012). The proof uses Gyárfás's intersecting-case theorem ($\tau\le 3$ for intersecting $4$-partite $4$-uniform families), a short projection lemma (four base-disjoint edges in an intersecting family force a two-element cover), and Kőnig's matching theorem.
Read the full abstract

2609.14281 · 13 Sep · 5 pages. Proves the (r,nu)=(4,2) case of Ryser's conjecture (Tuza's 1979 claim, recorded as open in DeBiasio et al. 2021). Every structural claim independently verified by exact computation (MILP + brute force). Proof found via four rounds of GPT-5.6 Pro; AI assistance disclosed in Methods

Analysis of PDEs

Lewy-Type Nondegeneracy for Gradient-Dependent Elliptic Equations and Global Gradient Diffeomorphisms on Convex Rings

We prove two complementary Lewy-type theorems for elliptic equations whose coefficients depend only on the gradient. First, let $\Omega\subset\mathbb{R}^3$ and let $u$ solve \[ a^{ij}(Du)u_{ij}=0, \] where $a$ is a smooth, symmetric, positive definite matrix field on an open set containing $Du(\Omega)$. We show that if the gradient map $Du$ is a local homeomorphism, then $\det D^2u$ never vanishes; hence $Du$ is a local $C^\infty$-diffeomorphism. Second, in every dimension, we consider capacitary solutions on convex rings. If the solution has no critical points and its level hypersurfaces are strictly convex, then ellipticity alone forces the Hessian to have one positive and $n-1$ negative eigenvalues, that is, inertia $(1,n-1)$. Consequently, the gradient is a global diffeomorphism onto a radially parametrized ring in gradient space. Both results apply to the $p$-Laplace and minimal surface equations. For the global minimal-surface result, existence of a smooth solution is assumed. Classical convex-ring results supply the noncriticality and strict level-set convexity needed in the global corollaries. Since the two coefficient matrices are real analytic on the relevant gradient ranges, the corresponding local and global gradient diffeomorphisms are real analytic.
Read the full abstract

Machine Learning

Biquaternionic Space with Complex-valued Attention for Temporal Knowledge Graph Completion

Temporal knowledge graph embedding (TKGE) models infer missing facts in knowledge graphs that evolve over time. Many existing models use a single geometric space, which can limit their ability to represent diverse relational patterns, or treat entity representations as static. We propose Biquaternionic Space with Complex-valued Attention (BSCA), a TKGE model that combines circular and hyperbolic rotations within a unified biquaternionic framework. A complex-valued attention mechanism adaptively fuses time-conditioned and relation-conditioned entity representations, allowing them to vary with temporal and relational context. Experiments on five benchmark datasets show competitive performance across datasets, with the largest improvement on GDELT: BSCA achieves an MRR of 52.1%, compared with 38.1% for the strongest baseline in our comparison.
Read the full abstract

Computer Vision

SpermYOLO: A Coordinated YOLO-Based Detector for Accurate and Efficient Sperm and Impurity Detection in Microscopic Images

Accurate sperm detection is essential for computer-assisted semen analysis, yet it remains challenging in microscopic images due to dense distributions, visually similar artifacts, and sperm-like impurities. In this paper, we propose SpermYOLO, a coordinated and compact YOLOv11-derived framework for joint sperm and impurity detection in microscopic images. SpermYOLO introduces four architectural improvements: C3k2-IDB for channel-wise discriminative feature extraction, D2SEM for spatial–spectral semantic enhancement, MFM for adaptive multi-scale feature fusion, and the DESD Head for detail-enhanced shared prediction. Experiments on the SVIA semen microscopic imaging benchmark show that SpermYOLO achieves 97.2% sperm AP and 75.4% impurity AP, outperforming generic detectors, dedicated sperm detection models, and improved YOLO variants. Compared with the baseline model, SpermYOLO improves sperm AP, impurity AP, $\mathrm{mAP}_{50}$, and $\mathrm{mAP}_{50:95}$ by 1.6, 10.0, 5.8, and 2.7 percentage points, respectively, while preserving a lightweight model scale. Cross-scene evaluation on the SDTB testicular-biopsy microscopy benchmark shows that SpermYOLO remains effective with extremely small sperm targets and complex tissue backgrounds, achieving the highest $\mathrm{mAP}_{50}$ and $\mathrm{mAP}_{50:95}$ of 74.8% and 31.2%, respectively. Ablation studies and qualitative analyses further support these improvements by demonstrating the contributions of the proposed modules and showing more focused feature response patterns than the baseline model. These findings suggest that SpermYOLO is an effective and efficient approach for sperm detection in challenging microscopic imaging scenarios.
Read the full abstract

All authors Shengqi Chen, Zilin Wang, Xingyu Pan, Wenting Yu, Pengchao Deng, Guohua Wu

General Relativity and Quantum Cosmology

Parameter Estimation of Ringdown Quasinormal Modes with Autoencoder

Ringdown gravitational waves from binary black hole mergers can be modeled as superpositions of quasinormal modes (QNMs), whose frequencies and excitation factors encode properties of the remnant Kerr black hole. Reliable extraction of multiple QNM components is challenging because of mode overlap and noise. We develop an autoencoder-based framework for multi-component QNM analysis, in which the latent space is trained to represent the physical parameters of individual modes, enabling waveform denoising and parameter estimation within a common framework. Using controlled model waveforms constructed as finite sums of Kerr QNMs with recently established high-precision frequencies and excitation factors, including their nontrivial spin dependence near resonant excitation, we assess the method across partitioned spin intervals. The model achieves good in-domain waveform reconstruction and parameter recovery for the two longest-lived components of eight-component input waveforms, while its performance degrades when the validation spins lie far outside the training range. In a selected spin interval, the framework also recovers the 32 parameters of an eight-component waveform with good overall agreement. These results demonstrate the feasibility of physics-informed autoencoder-based inference for a prescribed multi-component ringdown waveform family and motivate further tests with progressively more realistic signals.
Read the full abstract